When should a negative #COVID19 test be ignored? Let's say a doctor sees a patient w/ possible #COVID19. She highly suspects infection because patient has cough, chest pain, abnormal chest x-ray, fever, diarrhea & known exposure. Doc puts the chance at 90%. Next... (1/5)

... a throat swab is checked but comes back negative. Let's assume that throat swab is 90% sensitive & specific (in reality, it's lower than that, but let's assume a best case scenario). What is the chance this patient still has #COVID19? Read on... (2/5)

Turns out there's math for this. We use Bayes theorem. Good news is no need to memorize the equation. Just punch the numbers into this app (one of many): https://apps.apple.com/us/app/bayes-post-test-probability-calculator/id371920137. Pic shows scenario where pre-test likelihood is 90% & test is 90% sensitivity/specific. Next... (3/5)

... hit "run." If test comes back +ive then chance of #COVID19 rises to ~99%. But if it's negative, does that rule-out infection? No. Even with a negative test, still a 50% chance of infection. Need to test again, or test elsewhere (like in stool, in deep airway), or just...(4/5)

...assume patient has #COVID19 because clinical suspicion so high we ignore test result. In reality, real test is prob not 90% accurate. That's why clinical gestalt is so vital. The wisdom of clinicians should override test result unless & until there's 100% accurate test. (5/5)